Moderating Effects of Age and Gender on the Relationship Between Pediatric Obsessive–Compulsive Symptoms and Parental Accommodation

Youth and their caregivers were recruited as part of a larger study examining parenting practices and cognitive functioning in pediatric anxiety and related problems (e.g., OCS). Participants self-identified from flyers and website advertisements posted throughout Northeast Ohio. Children were included in the study if they were 7 through 17 years old. Though anxiety disorders were the focus of the larger study, no diagnosis was required to participate, and thus the study involved a mixed clinical/non-clinical sample (for similar mixed study populations in parental accommodation research, see [11]). Child exclusion criteria included (1) a diagnosis of major depressive disorder (assessed via structured diagnostic interview) and (2) a diagnosis of autism spectrum disorder (determined via parent-report). In total, 61 parents or guardians completed the Family Accommodation Scale—Anxiety (FASA), the Spence Children’s Anxiety Scale Parent Report–Obsessive–Compulsive Subscale (SCAS-P), and the Child Behavior Checklist (CBCL). Caregivers’ ages ranged from 31 to 67 (M = 44.25, SD = 7.0), and 90% of informants were female (i.e., child’s biological, adoptive, or foster mother). Children’s ages ranged from 7 to 17 (M = 11.92, SD = 2.3) and were 62% female. The majority of families were White/Caucasian, non-Hispanic/Latino, and had an annual household income greater than $60,001. No participants identified as transgender or nonbinary. See Tables 1 and 2 for complete demographic information, including parent/child race and ethnicity breakdown and child diagnostic information.

The FASA is a 13-item parent- and self-report measure of parental accommodation in anxiety disorders. Only the parent-report version was used for this study. Parents are asked to rate the frequency of accommodation on a 1 (Never) to 5 (Daily) point scale, with higher scores indicating more frequent accommodation. The FASA assesses parental accommodation in three domains: (1) parent participation in symptom-related behaviors (e.g., “How often did you reassure your child?”); (2) modification of functioning (e.g., “Have you modified your family routine because of your child’s symptoms?”); and (3) distress and consequences, rated on a 1–5 Likert scale ranging from No to Extreme (e.g., “Has your child become distressed when you have not provided assistance?”). The FASA is shown to have good internal consistency and convergent & discriminative validity among clinical populations [24].

The FASA was originally adapted for use from Calvocoressi and colleagues’ [6] measure of accommodation among relatives of individuals with OCD (FAS). Updated, self-report versions of the FAS for OCD remain in use (e.g., [13, 14, 30]), however, domains of accommodation measured across both the anxiety and OCD versions of the family accommodation scales are similar. A primary difference is the specific use of language such as “compulsions” or “rituals” in the FAS as opposed to more general language related to anxiety, avoidance, and reassurance in the FASA. Though OCS are the focus of the present analyses, the adapted FASA was opted for use in this larger project to capture a broader range of anxiety-related disorders in addition to OCS. Within this sample, the FASA demonstrated excellent internal consistency (α = 0.93).

The SCAS-P is a 38-item parent–report measure of anxiety symptoms across six domains: social phobia, separation anxiety, panic attack/agoraphobia, obsessive–compulsive disorder, generalized anxiety, and physical injury fears. Each item is rated on a 0 (Never) to 3 (Always) point scale, with higher scores indicating higher levels of anxiety. The SCAS-P has strong psychometric properties in children ages 8–12 (Spence et al., [36]) and acceptable psychometric properties in adolescents ages 13 and 14 [37]. In both age groups, the SCAS correlates strongly with established measures of child anxiety.

The obsessive–compulsive subscale was used as the primary outcome measure for this study. The subscale is composed of six questions assessing central features of OCD (e.g., “My child has to keep checking that he/she has done things right” or “My child can’t seem to get bad or silly thoughts out of their head”). SCAS-P: OCD subscale demonstrates good concurrent validity via strong correlations with other validated measures of OCD symptoms in clinical and community samples (e.g., Children’s Yale-Brown Obsessive–Compulsive Scale; [43]). Within this sample, the SCAS-P: OCD subscale demonstrated good internal consistency (α = 0.84).

The CBCL is a 112-item parent-report measure assessing broad adaptive, emotional, and behavioral functioning in youth ages 6–18. Parents are asked to rate behaviors on a 0 (Never True) to 2 (Often/Very True) point scale, with high scores indicating greater frequency of emotional/behavioral problems. The anxious-depressed and inattention subscales were used in the present study to control for anxiety, depression, and attention symptoms. The anxious-depressed subscale has been identified as a useful screening tool for anxiety and affective disorders in youth [2, 9]. The CBCL inattention subscale includes multiple facets of ADHD (e.g., inattention, hyperactivity, and impulsivity, e.g., “Can’t sit still”) and discriminates youth with and without ADHD [10, 41]. Within this sample, the CBCL anxious-depressed and inattention subscales demonstrate good internal consistency (α = 0.87, α = 0.89, respectively).

Two moderation analyses were conducted to determine whether age and gender moderate the relationship between parental accommodation and OCS severity. Child internalizing symptoms (i.e., CBCL: Anxious-Depressed subscale) and inattention (i.e., CBCL: Inattention subscale) were identified as possible covariates given higher rates of comorbidity among youth with OCD [17, 25]. Initial Pearson correlations were conducted between anxiety/depression, attention, and our primary outcome variable (SCAS-P: OCD subscale). After controlling for the number of comparisons using a Bonferroni correction, only the anxious-depressed subscale was significantly related to obsessive–compulsive symptoms (p < 0.001) and used as a covariate in the analyses. Basic assumptions of regression were tested prior to analyses. The SCAS-P: OCD subscale was not normally distributed. However, given our use of the PROCESS Macro that includes bootstrapping, a normal distribution is not critical, and the central limit theorem applies [12, 18].

A simple moderation analysis was conducted using SPSS’s PROCESS Macro [18] to evaluate whether child age moderated the relationship between parental accommodation and OCS, after controlling for child internalizing symptoms (i.e., CBCL: Anxious-Depressed subscale scores). The interaction between accommodation and age was significant (b = 0.05, p < 0.01), indicating that age moderated the relationship between accommodation and symptom severity (Table 4). The interaction was further probed by testing the conditional effects of accommodation at the 16th, 50th, and 84th age percentiles (Fig. 1). In middle (i.e., 50th percentile, M = 12 years of age) and older children (i.e., 84th percentile, M = 15 years of age), parental accommodation significantly predicted OCS. The relationship was stronger for the oldest group (Middle: b = 0.13, p < 0.01; Oldest: b = 0.27, p < 0.01). In younger children (i.e., 16th percentile, M = 9 years old), accommodation did not predict symptom severity (b = − 0.003, p = 0.94).

A simple moderation analysis was conducted to evaluate whether child gender moderated the relationship between parental accommodation and OCS, after controlling for child internalizing symptoms (i.e., CBCL: Anxious-Depressed subscale score). The interaction between accommodation and gender was significant (b = − 0.15, p = 0.01), indicating that gender moderated the relationship between accommodation and symptom severity (Table 4). The interaction was probed by testing the conditional effects of accommodation for each gender (Fig. 2). For boys, parental accommodation significantly predicted OCD symptom severity (b = 0.11, p = 0.02). For girls, accommodation did not predict symptom severity (b = − 0.04, p = 0.46).

A series of exploratory post-hoc t-tests were conducted to assess whether gender differences existed with respect to individual items on the Family Accommodation Scale – Anxiety (Table 5). Parents of boys reported significantly higher scores (M = 1.4, SD = 1.2) than parents of girls (M = 0.7, SD = 1.0) on the item “Has your child become distressed when you have not provided assistance?” [t(58) = 2.4, p = 0.02]. There was trending significance on the item “Has your child become angry/abusive when you have not provided assistance?” and on the Distress and Consequences subscale total, such that parents rated higher scores for boys than for girls (p = 0.067 and p = 0.073, respectively).

Results from our first set of analyses supported our hypothesis that parental accommodation may have a larger impact on adolescents’ OCS than on younger children’s OCS. Previously, researchers have noted that accommodation is more likely to occur in younger children [34],however, whether age-related differences exist regarding the relationship between accommodation and symptom severity has largely been unexplored within the OCS and parental accommodation literature. Though results are exploratory given the cross-sectional study design, it is possible that the relationship between parental accommodation and OCS mirror a relationship that occurs between the related concepts of parental psychological control (i.e., parents whose behaviors intrude upon their children’s independent thoughts and beliefs) and child anxiety. For example, an increasing correlation between parental psychological control and child anxiety severity has been shown as children transition into adolescence [20, 40, 44]. Furthermore, stronger pooled effect sizes have been demonstrated between parent psychological control and child anxiety in youth ages 13–15 (as opposed to youth ages 10–12 and 16–20) in at least one meta-analysis [40]. Though parental psychological control is a distinct construct from parental accommodation, these findings from previous studies are largely in line with our results suggesting the strongest relationship between parental accommodation and OCS in adolescence (i.e., M = 15 years). This suggests that youth may be particularly vulnerable to effects of parental overinvolvement in beliefs and routines as they transition from childhood into adolescence.

It is also possible that, more broadly, parental accommodation can be less detrimental for younger children’s OCS simply because younger children’s routines and behavior are more closely linked with their parents’ routines than that of an adolescent. Non-familial influences are limited in early/middle-aged children whose day-to-day schedule revolves heavily around their parents, as opposed to adolescents who engage in more autonomous behavior (e.g., choosing their own extracurricular activities, driving themselves, selecting their own peers/romantic partners; [20, 40]). Parental accommodation may be less disruptive to a child’s day when a parent–child routine is more closely linked. Consequently, it is possible that more internal and external conflict occurs when parents try to accommodate their teenager’s behavior to the same degree. This may lead to a worsening of OCS for a variety of reasons (e.g., worsening symptoms are a byproduct of family conflict, accommodation is logistically harder and less effective at reducing short-term distress, but youth still rely on it, etc.).

Girls’ anxiety symptoms are more heavily influenced by a parent’s own anxiety or their controlling/over-involved parenting practices [3, 31, 40]. This prior research guided our hypothesis that a similar relationship would exist with respect to the relationship between parental accommodation and OCS. Contrary to our prediction, we found that greater parental accommodation significantly predicted higher OCS in boys, but not in girls. One potential explanation for this contrary finding is that the experience of youths with anxiety does not extend to OCS. There is a notable and well-established increase in girls’ anxiety symptoms (as compared to boys’) during puberty, but the relationship between OCS and puberty is less clear cut [4]. Gender and parenting practices may interact differently with OCS than they do with anxiety symptoms.

Approximately 90% of the parent/guardian respondents within the current study were female (e.g., biological or adoptive mother). This imbalance is not uncommon in parent–child dyad research [48] but may nevertheless impact results. In general, mothers are more likely to accommodate or be involved in their children’s OCD-related rituals than fathers [28, 32]. There may also be gender differences both in the way mothers accommodate their children, and in the way mothers perceive and report their children’s OCS. More discrepancy exists between mother-son reports of symptoms than father-son reports of symptoms [19]. It is possible that mothers may misreport (or in this case, over-report) their sons’ OCS, particularly if they are more involved in compulsions (i.e., more accommodation). More research is needed to elucidate this specific relationship in the context of parental accommodation.

Finally, it is possible that gender differences exist regarding the types of OCS a child displays, or in the way children respond to accommodation or the lack thereof. This may subsequently impact the accommodation–symptom severity relationship and cycle. In an attempt to understand this unexpected gender-related finding, we conducted exploratory post-hoc t-tests to evaluate whether gender differences exist with respect to individual items on the FASA. Results indicated a significant gender difference on the item “Has your child become distressed when you have not provided assistance?” such that parents rated boys’ distress as greater than girls. Boys might exhibit more noticeable distress or externalizing behaviors when their parents do not accommodate, which in turn further reinforces accommodation and strengthens the relationship between accommodation and OCS. These findings are in line with the well-established notion that boys are often perceived as more aggressive than girls [8].

Our data did not allow for a gender comparison between types of OCS, but it is possible that this could have played a role as well (e.g., if boys have more compulsions that are disruptive to the family or seen as more disruptive than girls, accommodation may occur more frequently and more potently impact parent report of symptom severity). The role of gender in OCS presentation is not well understood, particularly in research with children, but some gender differences exist with respect to the types of obsessions and compulsions that persist [26]. Replication and extension of these findings are needed to clarify the relationship between gender, parental accommodation, and OCS. Importantly, our sample did not include any individuals who identified as non-binary or gender non-conforming. Future research should take into account differences between biological sex and gender identity when discussing sex or gender-related differences.